2032 - Real-Time Markerless Motion Tracking with Automatic Segmentation Method for Precise Magnetic Resonance Imaging-Guided Radiotherapy
Presenter(s)
S. Chen1, Z. Wang1, J. Dai2, G. Wu1, Y. Tang3, and J. Chen2; 1School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing, China, 2Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 3Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences (CAMS) and Peking Union Medical College (PUMC), Beijing, China
Purpose/Objective(s): Precise motion tracking is essential for the accurate delivery of MRI-guided radiotherapy (MRIgRT). This study aims to enhance motion tracking accuracy in MRIgRT by developing an automatic, real-time, markerless tracking approach that integrates an innovative tracking framework with automated segmentation.
Materials/Methods: We developed a novel motion tracking and segmentation approach for MRIgRT by combining the Enhanced Tracking-Learning-Detection (ETLD) algorithm with an Improved Chan-Vese (ICV) model, termed ETLD+ICV. The ETLD algorithm was optimized for MRI motion tracking, incorporating advanced image preprocessing, no-reference image quality evaluation, an improved median-flow tracker, and a refined detector with dynamic search region adjustments. The ICV model was applied for precise target volume delineation, iteratively refining segmented regions based on tracking results. The proposed method was evaluated using 3.5D MRI scans from 10 patients with liver metastases, analyzing a total of 106,000 MRI frames acquired from the Elekta Unity MR-LINAC system.
Results: Experimental results demonstrated that the ETLD method achieved sub-millimeter tracking errors, with a margin of less than 0.8 mm, alongside over 99% precision and 98% recall for all patients in the Beam Eye View (BEV) and Beam Path View (BPV) orientations. The ETLD+ICV approach yielded a Dice global score exceeding 82% for all patients, highlighting the method's robustness and ability to accurately cover the target volume.
Conclusion: This study presents an advanced automatic real-time markerless motion tracking method for MRIgRT, demonstrating superior performance compared to existing approaches. The proposed method ensures high accuracy in both motion tracking and segmentation, while also exhibiting improved adaptability to clinical requirements. This innovation positions the method as a valuable tool for enhancing the effectiveness and precision of radiotherapy treatments.